基于神经网络时用水量预测模型（终稿）.docVIP

PAGE  PAGE I 基于神经网络的时用水量预测模型 摘要 长期以来，精确预测时用水量是实现供水系统优化调度的关键，因此一直是城市供水部门比较关注的问题。但是由于时用水量影响因素较多，时用水量预测结果很难与实际用水量较好地吻合。本文根据时用水量序列的周期性、趋势性的特点，采用具有较强非线性映射能力和较好容错性特点的人工神经网络(ANN)法建立了时用水量预测模型。实例分析结果表明，该模型具有的较高预测精度，预测绝对相对误差的平均值仅为4.6％。与传统时用水量预测方法相比，该预测模型具有更好的预测精度。 关键词 神经网络 时用水量 时间序列 PAGE 31 WATER CONSUMPTION FORECAST MODEL BASED ON BP NEURAL NETWORK Name:Zhou Jianqiang Major subject: Water supply and drainage engineering Advisers:Chen Lei ABSTRACT Hourly water consumption forecast is a key step to realize optimal dispatch of water pipe network，so water works pay much attention to it. There are a lot of factors that influence the hourly water consumption, so it is difficult for the water consumption forecasting to fit the actual result well. According to the periodicity and trend of water demand series, an hourly water demand forecast model based on artificial neural network with the ability of high nonlinear mapping and good fault tolerance was developed. Case study shows that the model has high forecast performance, and the average of absolutely relative error is only 4.6%. Compared with traditional hourly water forecast methods, the forecast model based on artificial neural network has better prediction accuracy. Key words artificial neural network; hourly water consumption； time series 目 录  TOC \o 1-3 \h \z \u  HYPERLINK \l _Toc169871773 摘要 Ⅰ  HYPERLINK \l _Toc169871774 ABSTRACT Ⅱ  HYPERLINK \l _Toc169871775 第一章 绪论 1  HYPERLINK \l _Toc169871776 1.1 研究背景 1  HYPERLINK \l _Toc169871777 1.2 国内外时用水量预测方面的研究现状 2  HYPERLINK \l _Toc169871778 1.3 存在的主要问题和不足 7  HYPERLINK \l _Toc169871779 1.4 本文的主要工作 7  HYPERLINK \l _Toc169871780 1.5 本章小结 8  HYPERLINK \l _Toc169871781 第二章 神经网络介绍及其发展 9  HYPERLINK \l _Toc169871782 2.1 神经网络介绍 9  HYPERLINK \l _Toc169871783 2.1.1 神经网络的发展 9  HYPERLINK \l _Toc169871784 2.1.2 神经网络的应用及研究方向 11  HYPERLINK \l _Toc169871785 2.1.3 人工神经网络的特性 11  HYPER